Medical Image Indexing using Attributed Relational Graphs and Rectangular Trees

Main Article Content

Natarajan Meghanathan
Sujatha Sivakumar


Given a collection of medical images (like CT scans), we derive appropriate representations of their content and organize the images
together with representations in a multi-dimensional data structure so that we can search efficiently for images similar to an example image.
Image content is represented by Attributed Relational Graphs (ARGs) holding features of objects and relationships between objects. Our
proposed image indexing and similarity search methods rely on the assumption that a fixed number of “labelled†or “expected†objects (e.g.,
“heartâ€, “lungs†etc) are common in all images of a given application domain in addition to a variable number of “unexpected†or “unlabeledâ€
objects (e.g., “tumorâ€, “hematoma†etc). Our method can answer queries by example such as “find all X-rays that are similar to Smith’s X-ray.â€
The stored images are mapped to points in a multi-dimensional space and are indexed using state-of-the-art database methods (R-trees). The
proposed method has several desirable properties: (a) Database search is approximate so that all images up to a pre-specified degree of similarity
(tolerance) are retrieved; (b) it has no “false dismissals†(i.e. all images qualifying query selection criteria are retrieved) and (c) it is much faster
than sequential scanning for searching in the main memory and on the disk (i.e. by up to an order of magnitude); thus, scaling-up well for larger


Keywords: Indexing, Similarity Searching, Medical Images, R-tree, Attributed Relational Graphs


Download data is not yet available.

Article Details